GNOME Bugzilla – Bug 274329
Evolution crashes on load, with notification to d-bus enabled.
Last modified: 2013-09-10 14:03:47 UTC
Distribution: Debian 3.1 Package: Evolution Priority: Critical Version: unspecified Synopsis: Evolution crashes on load. Bugzilla-Product: Evolution Bugzilla-Component: Mailer Bugzilla-Version: unspecified BugBuddy-GnomeVersion: 2.0 (2.10.0) Description: Description of the crash: There is a couple of microseconds where I can do something, until it crashes. I think it might be related to viewing mails, but trying to switch to other mails has no effect. Steps to reproduce the crash: 1. 2. 3. Expected Results: How often does this happen? Every time I start Evolution. Additional Information: Debugging Information: Backtrace was generated from '/usr/bin/evolution-2.2' (no debugging symbols found) Using host libthread_db library "/lib/tls/libthread_db.so.1". (no debugging symbols found) `system-supplied DSO at 0xffffe000' has disappeared; keeping its symbols. (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) [Thread debugging using libthread_db enabled] [New Thread -1229616288 (LWP 16146)] [New Thread -1262416976 (LWP 16160)] [New Thread -1254028368 (LWP 16150)] [New Thread -1245639760 (LWP 16149)] [New Thread -1237251152 (LWP 16148)] (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) 0xb7aa9b61 in waitpid () from /lib/tls/libc.so.6
+ Trace 57618
Setting qa contact to the default for this product. This bug either had no qa contact or an invalid one.
Started evolution with -c calendar from a xterm, so got a bit more useful information. It was crashing on Generate a d-bus message when new mail arrives. Failed to connect to the D-BUS daemon: Unable to determine the address of the message bus 28657: arguments to dbus_connection_get_data() were incorrect, assertion "connection != NULL" failed in file dbus-connection.c line 4029. This is normally a bug in some application using the D-BUS library. 28657: arguments to dbus_connection_set_watch_functions() were incorrect, assertion "connection != NULL" failed in file dbus-connection.c line 3320. This is normally a bug in some application using the D-BUS library. ** ERROR **: Not enough memory to set up DBusConnection for use with GLib aborting...
might be the wrong version of the dbus api for evolution. i thought we patched that recently, but i can't see the patch in the code.
Created attachment 45062 [details] [review] fix that isn't applied yet
ok this patch is applied but fixes a different problem i think. there's a patch on the evolution patches list which looks atthis problem (null connection).
*** Bug 304093 has been marked as a duplicate of this bug. ***
*** Bug 303934 has been marked as a duplicate of this bug. ***
*** Bug 304713 has been marked as a duplicate of this bug. ***
Stack trace from Bug 304713 Backtrace was generated from '/opt/gnome/bin/evolution-2.2' Using host libthread_db library "/lib/tls/libthread_db.so.1". [Thread debugging using libthread_db enabled] [New Thread 1096133600 (LWP 31552)] [New Thread 1111608240 (LWP 31595)] [Thread debugging using libthread_db enabled] [New Thread 1096133600 (LWP 31552)] [New Thread 1111608240 (LWP 31595)] [Thread debugging using libthread_db enabled] [New Thread 1096133600 (LWP 31552)] [New Thread 1111608240 (LWP 31595)] [New Thread 1113709488 (LWP 31593)] [New Thread 1109375920 (LWP 31569)] [New Thread 1107274672 (LWP 31568)] [New Thread 1105083312 (LWP 31567)] 0xffffe410 in ?? ()
+ Trace 59900
Thread 1 (Thread 1096133600 (LWP 31552))
In the Bug I reported (304713), I have irritatingly concluded that the erasure of the file "~/.gconf/apps/evolution/mail/notify/%gconf.xml" would not help. Thats wrong. It does, if I erase it after I shutdown the gconfd and restart it. Moreover, I managed to run dbus-launch within my ~/.xsession right after gpg-agent and ssh-agent and before the window manager is run. That way a dbus-daemon-1 gets started, and re-checking that Option in Evolution does no longer lead to a crash. It looks like, that evolution does only require a check whether dbus-daemon is running, and if not, it shall not allow checking that option.
*** Bug 307632 has been marked as a duplicate of this bug. ***
*** Bug 307767 has been marked as a duplicate of this bug. ***
*** Bug 307909 has been marked as a duplicate of this bug. ***
*** Bug 308898 has been marked as a duplicate of this bug. ***
Here is a work-around that worked for me. Use the gnome configuration editor to DISABLE the following setting: /apps/evolution/mail/notify/gen_dbus_msg
Created attachment 48818 [details] [review] patch fixes related problems
Created attachment 48819 [details] [review] patch fixes related problems problem which i can see in the backtraces should be fixed in 2.3 I am attaching one more patch (for current CVS head) which should fix other bits of problems related to this.
committed the patch in CVS Head. Pretty sure its fixed on head. Closing it.
*** Bug 302955 has been marked as a duplicate of this bug. ***
*** Bug 310090 has been marked as a duplicate of this bug. ***
hmm, that patch is wrong anyway, the message CAN NEVER BE NULL. the bus can never be null either. you've tried to fix a bug on evolution-2.3 which was already fixed, the bug is in evolution 2.2 (and still is).
Created attachment 50256 [details] [review] fix against 2.2, which is completely different code to 2.4
looking at the d-bus code I found it was a good practice there to have a check everytime we call "dbus_message_new_signal" for the return value to be NULL and handle the case. I did not reproduce the crash though. Am still unsure why it can't be NULL.
*** Bug 320490 has been marked as a duplicate of this bug. ***
*** Bug 342595 has been marked as a duplicate of this bug. ***
*** Bug 351492 has been marked as a duplicate of this bug. ***
*** Bug 352715 has been marked as a duplicate of this bug. ***